from Hacker News

They Trusted ChatGPT to Plan Their Hike – and Ended Up Calling for Rescue

by speckx on 6/20/25, 10:50 PM with 4 comments

• by victorstanciu on 6/21/25, 5:52 AM

  > The pair, wearing only flat-soled sneakers [...]

  I appreciate a good "AI" hate session as much as the next man, but these people would have had to be rescued no matter what tool they used to plan their trip, as long as the one tool they should have used--their brain--was apparently already on vacation ahead of them.

• by joshstrange on 6/21/25, 6:53 PM

  LLMs are not the second coming but to blame this on ChatGPT seems far-fetched. This isn’t quite “Michael Scott driving his car into a lake” but it’s pretty darn close.

  While I’m glad these people are ok, I can’t help but think more people need to feel stronger consequences for their mistakes. Hopefully they got slapped with a large fine for wasting emergency service resources.

• by mindslight on 6/21/25, 3:48 AM

  Is there some way to file a reverse class action against everyone calling these tools "Artificial Intelligence" ? The information spit out by LLMs is like two quality grades up from asking 4chan, yet that doesn't stop the pushers from framing them as some omniscient gods.

  I think the main thing the trend shows is just how bad people are at checking information coming at them. I'd estimate that around 75% of the questions I ask ChatGPT, it spits out blatantly wrong answers for. And even clarification/refining/outright correcting is still unable to fix like a third of those. It's really good at generating narratives that sound polished though.